advancing instead of speeding up?!?!?
intel chopped out their barrel shifters because they couldnt run at 1.35+ghz! they cut out a full speed full size predecode cache and put in a 16kbyte (yes 16kbyt) post decode cache, when a post decode cache really needs to be larger since a single x86 intruction encodes to between 2 and 3 risc operations for the core. and they cut out one of the decode units. why? so that their core is smaller so they can pump the clock speed up.
basically intel chopped their cpu to bits so that they could speed their clock up since most people think of clock speed as the beall and endall of cpu performance.
if you are using code that is optimized just for the p4 it is faster than generic code. but the same goes for athlon, and generic code runs way faster on athlon than p4. especially stuff with lots of jumps wich screw up p4s decoder and cache setup or calculated address stuff that uses the hackedup shifter setup. :-o
oh just incase you were wondering i did read all the tech docs (white papers) from intel and amd.
